The Task
Script stock prices of top 40 companies in Singapore by market capitalization between 1st January 2020 - 31st December 2021 by using tidyquant R package.
Using either calender heatmap or horizon graph, prepare a data visualization showing the historical stock prices by the top 40 companies in Singapore by market capitalization.
Installing and Loading the Required Libraries
The code chunk below is to install the packages needed.
- tidyquant is to get quantitative data in tibble format.
- tidyverse is to clean data.
- ggHoriPlot is to build horizon plots in ggplot2.
- ggTimeSeries is to build time series visualizations Using the Grammar of Graphics.
- kableExtra to create cleaner looking tables for display.
packages = c('tidyquant', 'tidyverse', 'ggHoriPlot', 'ggTimeSeries', 'kableExtra', 'ggthemes', 'plotly')
for(p in packages){library
if(!require(p, character.only = T)){
install.packages(p)
}
library(p, character.only = T)
}
Data Scraping
The symbols of the top 40 companies in Singapore by market capitalization was downloaded as companiesmarketcap.com - Largest companies in Singapore by market capitalization.cs. This csv data was imported using below code.
- read_csv() is a function of readr package to import csv file.
symbol_table <- read_csv("data/companiesmarketcap.com - Largest companies in Singapore by market capitalization.csv") %>%
rename(symbol = Symbol)
Below code chunk is to script the stocks data from Yahoo Finance.
- tq_get() is a one-stop shop for getting web-based financial data in a “tidy” data frame format. Get data for daily stock prices (historical), key statistics (real-time), key ratios (historical), financial statements, dividends, splits, economic data from the FRED, FOREX rates from Oanda.
- tq_mutate() is used to add a column to the data frame.
- tq_transmute() is used to return a new data frame which is necessary for periodicity changes.
- tq_performance() converts investment returns into performance metrics.
- tq_portfolio() aggregates a group (or multiple groups) of asset returns into one or more portfolios.
from_date = "2020-01-01"
to_date = "2021-12-01"
stock_selected = c(symbol_table$symbol[1:40])
Top40_Name = c(symbol_table$Name[1:40])
stock_data_daily = tq_get(stock_selected,
get = "stock.prices",
from = from_date,
to = to_date)
kable(head(stock_data_daily)) %>%
kable_styling()
| symbol | date | open | high | low | close | volume | adjusted |
|---|---|---|---|---|---|---|---|
| SE | 2020-01-02 | 41.00 | 41.190 | 39.460 | 40.04 | 5857000 | 40.04 |
| SE | 2020-01-03 | 39.44 | 40.650 | 39.400 | 40.49 | 5237800 | 40.49 |
| SE | 2020-01-06 | 40.07 | 41.030 | 40.029 | 40.48 | 4049200 | 40.48 |
| SE | 2020-01-07 | 40.50 | 41.800 | 40.350 | 41.01 | 3947600 | 41.01 |
| SE | 2020-01-08 | 41.00 | 41.389 | 40.055 | 40.16 | 3200800 | 40.16 |
| SE | 2020-01-09 | 40.79 | 40.900 | 40.040 | 40.53 | 3129700 | 40.53 |
Let’s now join ‘stock_data_daily’ with ‘symbol’ to include the company name in the data tibble.
stock_data_daily1 <- stock_data_daily %>%
left_join(symbol_table,
by = "symbol")
Next, we want to order the stock in descending order by the market capitalization value using order().
stock_data_daily1 <- stock_data_daily1[order(-stock_data_daily1$marketcap),]
Alternative method is as below:
- mutate() adds new variables and preserves existing ones.
- factor() is used to encode a vector as a factor (the terms ‘category’ and ‘enumerated type’ are also used for factors). If argument ordered is TRUE, the factor levels are assumed to be ordered.
- arrange() is to arrange rows by column values.
stock_data_daily1 <- stock_data_daily1 %>%
mutate(Name = factor(Name, levels = Top40_Name)) %>%
arrange(Name)
kable(head(stock_data_daily1)) %>%
kable_styling()
| symbol | date | open | high | low | close | volume | adjusted | Rank | Name | marketcap | price (USD) | country |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SE | 2020-01-02 | 41.00 | 41.190 | 39.460 | 40.04 | 5857000 | 40.04 | 1 | Sea (Garena) | 76666667008 | 137.46 | Singapore |
| SE | 2020-01-03 | 39.44 | 40.650 | 39.400 | 40.49 | 5237800 | 40.49 | 1 | Sea (Garena) | 76666667008 | 137.46 | Singapore |
| SE | 2020-01-06 | 40.07 | 41.030 | 40.029 | 40.48 | 4049200 | 40.48 | 1 | Sea (Garena) | 76666667008 | 137.46 | Singapore |
| SE | 2020-01-07 | 40.50 | 41.800 | 40.350 | 41.01 | 3947600 | 41.01 | 1 | Sea (Garena) | 76666667008 | 137.46 | Singapore |
| SE | 2020-01-08 | 41.00 | 41.389 | 40.055 | 40.16 | 3200800 | 40.16 | 1 | Sea (Garena) | 76666667008 | 137.46 | Singapore |
| SE | 2020-01-09 | 40.79 | 40.900 | 40.040 | 40.53 | 3129700 | 40.53 | 1 | Sea (Garena) | 76666667008 | 137.46 | Singapore |
Now, our dataset ‘stock_data_daily1’ is ready for visualization.
Data Visualization
Calender Heatmap
A calendar heatmap is a great way to visualise daily data. Its structure makes it easy to detect weekly, monthly, or seasonal patterns.
We firstly create a function to make a calendar heatmap and set Name as the variable, so that we don’t need to repeat the steps when plotting for all the 40 companies. Later, we will loop through all the values in Name column to create the corresponding plots.
- Plot title is set according to the company name. Unit is added after the company name.
- Low value is set to use red color and high value to use green color
- Theme is customized using theme() function: all the text are bold. Gridlines are grey. Background has no fill. Border is black.
calendar_heatmap <- function(df, var) {
ggplot_calendar_heatmap(
df %>% filter(Name ==var),
'date',
'close',
dayBorderSize = 0.25,
dayBorderColour = "black",
monthBorderSize = 1,
monthBorderColour = "black",
monthBorderLineEnd = "round") +
xlab('') +
ylab('') +
labs(title = paste(var, "Stock Close Price (SGD/share)"))+
scale_fill_continuous(low = "red", high = "green")+
theme(plot.title = element_text(size = rel(1), hjust=0.5, face = "bold"),
panel.background = element_rect(fill = NA),
panel.grid.major = element_line(colour = "grey90"),
panel.border = element_rect(linetype = "solid", fill = NA),
strip.background = element_rect(colour = "black"),
strip.text.x = element_text(colour = "black", size = rel(1), face = "bold"),
legend.title = element_text(face = "bold"),
axis.text = element_text(colour = "black", size = rel(0.8), face = "bold"))
}
Next, we loop through Name var, create plots and store them in a list using purrr::map(). Then, we display all plots with purrr::walk().
- The map functions transform their input by applying a function to each element of a list or atomic vector and returning an object of the same length as the input.
plot_list <- unique(stock_data_daily1$Name) %>%
purrr::set_names() %>%
purrr::map( ~ calendar_heatmap(stock_data_daily1, .x))
purrr::walk(plot_list, print)
Below is to explore if above calendar heatmap works with ggplotly() to make it interactive. We can see that this interactive plot works.
p <- ggplot_calendar_heatmap(
stock_data_daily1 %>% filter(Name == 'DBS'),
'date',
'close',
dayBorderSize = 0.25,
dayBorderColour = "black",
monthBorderSize = 1,
monthBorderColour = "black",
monthBorderLineEnd = "round") +
xlab('') +
ylab('') +
labs(title = paste('DBS', "Stock Close Price (SGD/share)"))+
scale_fill_continuous(low = "red", high = "green")+
theme(plot.title = element_text(size = 8, hjust=0.5, face = "bold"),
panel.background = element_rect(fill = NA),
panel.grid.major = element_line(colour = "grey90"),
panel.border = element_rect(linetype = "solid", fill = NA),
strip.background = element_rect(colour = "black"),
strip.text.x = element_text(colour = "black", size = 8, face = "bold"),
legend.title = element_text(face = "bold"),
axis.text = element_text(colour = "black", size = 8, face = "bold"))
ggplotly(p, height = 250, width=900)
Horizon Graph
Horizon Graph compact the area chart by slicing it horizontally, and then then shifting the slices to baseline zero. It’s like a combo area chart and heatmap. The figure below shows a horizon plot using the ggHoriplot package.
- Origin is set at mean value.
- Horizon scale is set at 6, which means cut the area chart into 6 pieces.
- Color palette chose is ‘RdBu’. Red color means values below mean value. Blue color means values above mean value.
- Theme is customized using theme() function for better visualization.
- X axis label is adjusted to show both the month and the year.
- Title and subtitle are added for explanation.
ggplot() +
geom_horizon(aes(date, close),
data = stock_data_daily1,
origin = 'mean',
horizonscale = 6) +
scale_fill_hcl(palette = 'RdBu') +
facet_grid(Name~.)+
theme_few() +
theme(panel.spacing.y=unit(0, "lines"),
plot.title = element_text(size=13, face = "bold"),
plot.subtitle = element_text(size=9, face = "bold"),
strip.text.y = element_text(size = 9, angle = 0, hjust = 0, face = "bold"),
axis.text.y = element_blank(),
axis.text.x = element_text(size=9, angle = 90, face = "bold"),
axis.title.y = element_blank(),
axis.title.x = element_blank(),
axis.ticks.y = element_blank(),
panel.border = element_blank(),
legend.position="none") +
scale_x_date(expand=c(0,0),
date_breaks = "1 month",
date_labels = "%b-%y",
limit=c(as.Date("2020-01-01"),as.Date("2021-12-31"))) +
ggtitle('Daily Stock Price of Top 40 Companies in Singapore (SGD/share), 2020 to 2021',
'Ranked by Market Capitalisation')
Below is to explore if above horizon graph works with ggplotly() wrapper to make it interactive.
H <- ggplot() +
geom_horizon(aes(date, close),
data = stock_data_daily1,
origin = 'mean',
horizonscale = 6) +
scale_fill_hcl(palette = 'RdBu') +
facet_grid(Name~.)+
theme_few() +
theme(panel.spacing.y=unit(0, "lines"),
plot.title = element_text(size=13, face = "bold"),
plot.subtitle = element_text(size=9, face = "bold"),
strip.text.y = element_text(size = 9, angle = 0, hjust = 0, face = "bold"),
axis.text.y = element_blank(),
axis.text.x = element_text(size=9, angle = 90, face = "bold"),
axis.title.y = element_blank(),
axis.title.x = element_blank(),
axis.ticks.y = element_blank(),
panel.border = element_blank(),
legend.position="none") +
scale_x_date(expand=c(0,0),
date_breaks = "1 month",
date_labels = "%b-%y",
limit=c(as.Date("2020-01-01"),as.Date("2021-12-31"))) +
ggtitle('Daily Stock Price of Top 40 Companies in Singapore (SGD/share), 2020 to 2021',
'Ranked by Market Capitalisation')
ggplotly(H)
We can see that this interactive plot doesn’t work. Only x and y axis and title are displayed.
Conclusions
- In general, most stock prices performed better in 2021 than 2020.
- There is a clear stock price drop in Feb/Mar 2020 when Covid-19 started to spread in Singapore.
- Sea (Garena) stock price has been steadily increasing since Jan 2020, which seems to be not impacted by Covid-19.
- Travel, hotel, shopping, real estate, construction and transportation industries haven’t yet recovered to the level of before Covid-19. Some of these companies are Singapore Airlines, CapitaLand Mall, Genting Singapore, Jardine singapore, Suntect, SATS and ComfortDelGro, City Development, United Industrial, ST Engineering, Keppel, Sembcorp and so on.
- Banking industry fully recovered from Covid-19 impact and stock price increased yoy. Examples are DBS, OCBC, and UOB.
- Trust companies stock price only suffered mildly from Covid-19. Netlink Trust and Mapletree Logistics Trust are the good performers.
Comparision between Calendar Heatmap and Horizon Graph
- Calendar heatmap shows more details in terms of the weekly cycle than the horizon graph.
- Calendar heatmap can show the stock price in terms of the actual values. Horizon graph shows the deviation from origin or a benchmark value. The darker the color is, the bigger the deviation is.
- When comparing the differences across many different companies, horizon graph is the better choice because it can accommodate all the data into a single compact graph. Whereas, multiple calendar heatmaps are needed in order to visualize all the 40 companies. facet_wrap() and facet_grid() allow multiple subplots in a single plot. However, 40 subplots into one plot are difficult to read.
- ggplot_calendar_heatmap() works with ggplotly() wrapper to become interactive. geom_horizon() doesn’t work with ggplotly() wrapper.
Reference
Using R to Scrape Financial Dataset
ggHoriPlot: build horizon plots in ggplot2